Cloud PMS Support Boundary — Who Owns What

Why define the boundary?

Because an incident at 23:00 does not care about the contract. If the boundary is unclear, the desk is passed between the PMS vendor, the network provider and the lock vendor while the guest waits.

What does the vendor own?

The application, the database environment, hosting, backups and patching. It also owns the interfaces it delivers and the support path for its side.

What does the property own?

The network, the printers, the terminals, the locks, the local procedures and the people. The property also owns the data quality and the operating decisions.

Who owns an interface?

The vendor that delivers the interface owns its side; the third-party vendor owns its system. The contract should name the coordinator for a cross-vendor failure.

How should an incident be coordinated?

One named coordinator, a single incident channel and a written escalation path. The property should not have to decide which vendor to call while the queue grows.

What should the contract say?

The layers, the owners, the response targets, the escalation path and the fallback procedures. A support matrix with names is more useful than a paragraph of goodwill.

How does the boundary work across a group?

A group adds a second layer: the property, the central IT team and the vendor. The property owns the local network and devices; the centre owns the standard, the identity provider and the reporting definition; the vendor owns the application environment and the interface it delivers.

Write the matrix once and reuse it at every property. The names change; the layers and escalation path should not. A group should also decide whether first-line support sits at the property, the centre or the vendor, because that decision drives the response time the guest experiences. The goal is a single path the desk can follow at 23:00, even when the first contact is unsure who owns the layer.

Test the matrix with a drill that crosses layers: a network outage during arrivals, a payment terminal failure at checkout and a lock failure at 23:00. Record who coordinates, how long each handoff takes, which fallback procedure was used and what the guest was told. The drill is more useful than a paragraph in the contract because it exposes the gaps between vendors.

• Name one incident coordinator per property and one at the centre.
• Publish the escalation path with roles, not just phone numbers.
• Keep a local fallback procedure for arrivals, payments and keys.
• Review the matrix after every cross-vendor incident.